Two interesting papers on diabetes in this week's NEJM.
In the USA, Type 1 diabetes is increasing slowly (1% per annum) among people aged under 20 years. Type 2 diabetes is increasing much faster (7% per annum) but from a much lower baseline. There are clear ethnic differences in incidence. This is large study with strong ascertainment and a requirement for physician diagnosis. I guess the increase in Type 2 is due to obesity, but the cause of Type 1 diabetes remains a mystery.
Note the use of ARMA to analyse the time trend. (I said 'note' not 'understand' :) )
Over in Sweden they've looked at mortality in people over 18 years, using the impressive nationwide register. Cox proportional hazards for this analysis, and a control matched to each case to reduce confounding by other improvements in general mortality. Huge declines over the past 15 years for both Type 1 and Type 2. The authors seem to think that this is because of better control of cardiovascular risk factors as well as better management of diabetes.
Figure 2 looks at Type 2 diabetes. Three points to make: (1) Death rates are higher than for the matched controls; (2) for CVD deaths, the decline in diabetics parallels the decline in the controls; but (3) for all cause deaths, the decline in diabetics seems to have plateaued over the past 10 years.